Insights

Ideas tested in public.

One idea from the book each month, with a question at the end. Your replies go back into the manuscript. The latest pieces are below.

Join the launch list Subscribe on Substack

Writing

13 November 2026 · Trusted capability AI Certified. Capability Not Found. Licences bought, staff certified, pilots launched: none of these move the numbers on their own. Capability does. OUTCOMES sets out the eight elements that turn AI investment into capability, and a board can find its binding constraint in one meeting. Read the piece → 6 November 2026 · Agentic insider The Debate Is Polarised. The Evidence Is Not. The AI slowdown debate split into three camps arguing from the same incident. Here is the ledger the debate should be having: 45 documented incidents of AI agents crossing a boundary, coded, graded, and turned into what a board should do this quarter. Read the piece → 26 October 2026 · Orientation debt The Better Aircraft Lost In 1953 the better aircraft lost. John Boyd spent forty years working out why, and the answer is the most misunderstood idea in leadership, and a warning for any board buying AI. Read the piece → 26 October 2026 · Board briefing Ten Questions Every UK Board Should Ask About AI Every UK board is now accountable for AI decisions it did not build. Ten questions, the evidence to ask for and the UK law behind each, with a way to score the answers on evidence rather than confidence. Read the piece → 1 October 2026 · Question 06 of ten Who signed for the model? Every AI model a board depends on needs one accountable name and a tested way to stop it. An Army lesson in ownership, applied to the boardroom. Read the piece →
Talks

Science wins the argument. Art wins the decision.

Keynotes and board briefings on leadership in a world of AI. Recent talks on YouTube; invitations by the briefing link.

Watch on YouTube Invite me to speak